Production Safety (AI)

Location: London, United Kingdom; Mountain View and New York City, United States; Zurich, Switzerland

Base salary: USD $119,800–$234,700 per year in the U.S.; USD $160,200–$261,000 per year in the San Francisco Bay Area and New York City metropolitan area.

Company

Microsoft AI develops AI systems and products, including frontier models, product engineering, and responsible AI technologies.

What you will do

• Design, build, deploy, and operate safety-critical services in production inference paths for text, multimodal, and agentic AI systems.
• Integrate model-based classifiers, policy engines, and guardrails with model APIs and serving platforms.
• Build containerized Kubernetes services with CI/CD, staged regional rollouts, rollback, and failover mechanisms.
• Define service-level objectives for availability, latency, throughput, correctness, and cost, including capacity planning and autoscaling.
• Develop monitoring and debugging capabilities for service health, safety decisions, training pipelines, and evaluation runs.
• Validate launches and contribute to on-call support, incident response, root-cause analysis, and reusable operational standards.

Requirements

• Bachelor’s degree in Computer Science, Engineering, a related technical field, or equivalent practical experience.
• Strong software engineering skills in one or more production languages such as C++, C#, Java, Go, Rust, or Python.
• Experience designing, deploying, and operating distributed services or large-scale production systems.
• Experience with containerized workloads, Kubernetes, and automated CI/CD pipelines.
• Knowledge of observability, capacity planning, failure isolation, graceful degradation, incident response, security, privacy, and system failure modes.
• Ability to collaborate across engineering, machine learning, product, and policy teams and communicate system tradeoffs clearly.

Nice to have

• Experience with online inference platforms, model serving, API gateways, policy enforcement systems, or latency-sensitive infrastructure.
• Experience operating machine learning models, training pipelines, or evaluation systems in production.
• Experience with Azure technologies, observability tooling, or capacity planning.
• Familiarity with AI safety, trust and safety, abuse prevention, content moderation, security engineering, privacy, or responsible AI.

Culture & Benefits

• Work with interdisciplinary teams across frontier model development, product engineering, responsible AI, security, and production inference.
• Collaborate with Microsoft product organizations, external partners, and customers.
• Some roles may be eligible for benefits and additional compensation.
• Applications are accepted on an ongoing basis until the position is filled, with the posting open for at least five days.